Scribe

Team

Age: under 18

Inspiration

We got our inspiration for our project, Scribe, from the increasing amount of word to voice texting that occurs around social media platforms. We believe that it’s an extremely efficient way to get messages across, so as high school students, we decided to implement this technique into our studying.

What it does

Scribe is a program written to enhance the learning process and other affairs by turning uploaded audio recordings into text. Although our main vision was for Scribe to be used by students in order to supplement their education. However, our program can also be used in other scenarios, such as transcribing court proceedings, helping hearing impaired people visualize speech, and allowing people to render long speeches into text.

How we built it

Scribe was built using Jupyter Notebook and the AssemblyAI API. We used Tkinter in order to request for the user to upload a local file from their computer that they would like to transcribe. The local file is then uploaded to the AssemblyAI backend server where it is given an unique identifier and queried for translation. Finally, once the file finishes translation, the JSON text is retrieved from the transcription endpoint within AssemblyAI and saved as a text file within the user's local files. The text is then displayed as well within the program itself.

Challenges we ran into

We hoped to create a UI in order to go along with our foundational program. However, many issues with Android Studio arose, such as trying to call Python script within a Java program. Sharing Android Studio files between computers also brought up complications due to differences in file path. We hope to use another software in the future to develop our application.

Accomplishments that we're proud of

A working program! We also troubleshooted and brainstormed as a team to solve our problems.

What we learned

We learned that integrating a Python script into a Java program is very difficult given our level of programming skill. In addition, we gained experience working with Jupyter Notebook and using Tkinter to get local files from a user. AssemblyAI was also new to us, as was speech to text in general. We were really proud of being able to grasp a foreign concept like that in the little amount of time we had.

What's next for Scribe

The immediate next step for Scribe would be to create a UI and hopefully an app in the future where the user can upload and transcribe on the spot. Another future addition would be to make Scribe into an all encompassing notes app where the user can type, upload, and live record notes. We also hope to add accessibility features such as unique typefaces for people with dyslexia, text to speech audio, and varying contrasting color schemes.